This module contains definitions and metadata for exceptional language codes: those that consist of two- or three-letter sequences separated by hyphens. See Wiktionary:Languages for more information.
This module must not be used directly in other modules or templates. The data should be accessed through Module:languages.
Every entry in the table must contain the following indexed fields:
- The "canonical" name of the language. This is the name that is used in Wiktionary entries and category names.
- The Wikidata item id (Q number) for the language. This replaces the older
wikipedia_articleproperty. Set to
nilif not known/present.
- The code for the family that the language belongs to. See Wiktionary:Families.
sort_key are used for text substitution; they replace or remove certain characters or sets of characters. They both work similarly, and are optional. They can both be tables, and
sort_key can be the name of a module that takes an entry name and generates a sortkey (which is used to sort the entry on category pages).
sort_key is the name of a module, the module must contain a sortkey-generating function that is named
makeSortKey. This function must take the arguments
text, lang, sc, where
text is the page name (or other text in the language),
lang is the language code (not the language object), and
sc is the script code (not the script object). The returned value should always be a string, or there will be a module error in the
If either one is a table, it must contain two tables inside it: one named
from and one named
to. These two tables are organised pairwise: each element in
from is a pattern to identify which characters in the term to replace, while the corresponding element in
to defines what to replace them with.
If the replacement is not present or if it is
nil, it defaults to an empty replacement, meaning that the matching characters are removed altogether. This means that the
from list can be longer than the
to list, and an empty replacement will be assumed for any elements in
from that have no counterpart in
The tables can contain literal characters, or the patterns (a type of regular expressions) that are used by the standard Scribunto
mw.ustring.gsub function. See the Scribunto reference manual for more information.
At the top of the module, there is a list of combining characters with names. These are provided for convenience and readability, as combining characters generally do not display properly inside the module code (although they do not affect the actual operation of the module).
- Defines replacements to create the entry name from the displayed form of a term. This can be used to remove certain diacritical marks according to the customs or standard practice of the language. For example, it is used to remove accent marks from Russian words (
русский), or macrons from Latin or Old English words (
os), as these are not used in the normal written form of these languages. This is used by
- Defines replacements to create a category sort key from the page name. The purpose is to remove any characters that are ignored in sorting, and to replace similar characters with identical ones if the sorting rules for that language do not distinguish them. For example, in German, the characters "ä" and "a" are considered equivalent for sorting, and are both treated as "a". The page name is converted to lowercase before applying the replacements, so you should not add uppercase letters to the "from" lists. This is used by
These are other optional values:
- A table of all the names that this language is known by, other than the canonical name. The list should include not only synonyms for the language, but also names that refer to language varieties that are subsumed under the same grouping. For example, while "Flemish" is not synonymous with "Dutch", Flemish is considered a part of Dutch, so the name is listed there.
- The type of language (which affects how it is handled on Wiktionary). Possible values are:
regular- This value is the default, so it doesn't need to be specified. It indicates that the is attested according to WT:CFI and therefore permitted in the main namespace. There may also be reconstructed terms for the language, which are placed in the Reconstruction namespace and must be prefixed with * to indicate a reconstruction.
reconstructed- This language is not attested according to CFI, and therefore is allowed only in the Reconstruction namespace. All terms in this language are reconstructed, and must be prefixed with *.
appendix-constructed- This language is attested but does not meet the additional requirements set out for constructed languages (WT:CFI#Constructed languages). Its entries must therefore be in the Appendix namespace, but they are not reconstructed and therefore should not have * prefixed in links.
- A list of script codes, see Wiktionary:Scripts. These represent all the scripts (writing systems) that this language uses in the real world, as well as the ones that Wiktionary uses. The scripts that are used most often on Wiktionary should be first in the list, as this will speed up script detection.
- Many templates and modules detect the script of text in a particular language using the
findBestScriptfunction in Module:scripts. This function goes down the list of scripts and counts how many characters in the text belong to each script. If all the characters belong to one script, that script will be returned; otherwise, the script with the most characters will be returned. Thus, script detection will be faster if the most frequently used scripts are first in the list.
- The name of a module that is used to generate transliterations of terms, without the Module: prefix. This module must export a function named
trthat is defined as follows:
tr(text, lang, sc)
- The three parameters are the text to be transliterated, the language code, and the script code. The function can ignore the language and script codes, but they are provided for cases when a language has more than one script, or when a single function is used to transliterate multiple languages sharing the same script.
- A table listing the language codes of the direct ancestors of this language. For example, the ancestor of English is listed as
ang(Old English, the ancestor of Middle English),
gem-pro(Proto-Germanic, the ancestor of Old English), and
ine-pro(Proto-Indo-European, the ancestor of Proto-Germanic) are not listed.
- For most languages, only one ancestor code should be given, but multiple ancestors can be listed for pidgins, creoles and mixed languages.
- The ancestor language table should not be included if the language's direct ancestor is the proto-language of the family to which the language belongs. In such a case, if the family code has been provided, Module:languages will automatically add the proto-language as the language's ancestor. For example, Proto-Germanic (
gem-pro) belongs to the Indo-European (
ine) family, and its direct ancestor is Proto-Indo-European (
ine-pro). Because Proto-Indo-European is the proto-language of the Indo-European languages, Proto-Germanic does not need an
ancestorstable; Proto-Indo-European will be automatically returned as its ancestor by the
- A table listing the Wikimedia language codes that this language maps to. This is used to translate Wiktionary codes to Wikimedia codes, which are usually the same but there are a few languages where it is different. The language codes must be valid Wikimedia codes (as determined by the wiki software), and if they are not defined in one of the language data modules, they must be defined in Module:wikimedia languages/data.
- The name of the Wikipedia article for the language. Should normally only be supplied if the Wikidata id cannot be used.